Time-series data is mostly based on observations around us and many physical models are based on differential equations.
However, accounting for non-ideal behaviour is difficult and using approaches such as regression does not capture the underlying dynamics.
Neural Differential Equations are an upcoming class of tools that allow for enhanced predictions by capturing the time-series not as a whole but by "fitting" the underlying dynamics.
Also included is the Stochastic Differential Equation model as well to capture dynamics in noisy situations.
NODEFit.jl is the faster Julia-based cousin that can be utilized if you are willing to live with the compile times.
GPU support is part of the PyTorch routines itself.
Time-series data is mostly based on observations around us and many physical models are based on differential equations.
However, accounting for non-ideal behaviour is difficult and using approaches such as regression does not capture the underlying dynamics.
Neural Differential Equations are an upcoming class of tools that allow for enhanced predictions by capturing the time-series not as a whole but by "fitting" the underlying dynamics.
Also included is the Stochastic Differential Equation model as well to capture dynamics in noisy situations.
NODEFit.jl is the faster Julia-based cousin that can be utilized if you are willing to live with the compile times.
GPU support is part of the PyTorch routines itself.